Phylodynamics of HIV-1 subtype C epidemic in east Africa

Abstract

The HIV-1 subtype C accounts for an important fraction of HIV infections in east Africa, but little is known about the genetic characteristics and evolutionary history of this epidemic. Here we reconstruct the origin and spatiotemporal dynamics of the major HIV-1 subtype C clades circulating in east Africa. A large number (n = 1,981) of subtype C pol sequences were retrieved from public databases to explore relationships between strains from the east, southern and central African regions. Maximum-likelihood phylogenetic analysis of those sequences revealed that most (>70%) strains from east Africa segregated in a single regional-specific monophyletic group, here called C(EA). A second major Ethiopian subtype C lineage and a large collection of minor Kenyan and Tanzanian subtype C clades of southern African origin were also detected. A bayesian coalescent-based method was then used to reconstruct evolutionary parameters and migration pathways of the C(EA) African lineage. This analysis indicates that the C(EA) clade most probably originated in Burundi around the early 1960s, and later spread to Ethiopia, Kenya, Tanzania and Uganda, giving rise to major country-specific monophyletic sub-clusters between the early 1970s and early 1980s. The results presented here demonstrate that a substantial proportion of subtype C infections in east Africa resulted from dissemination of a single HIV local variant, probably originated in Burundi during the 1960s. Burundi was the most important hub of dissemination of that subtype C clade in east Africa, fueling the origin of new local epidemics in Ethiopia, Kenya, Tanzania and Uganda. Subtype C lineages of southern African origin have also been introduced in east Africa, but seem to have had a much more restricted spread.

Figure 1. Maximum likelihood phylogenetic tree based on 1,981 HIV-1 subtype C pol (∼1,000 pb) sequences.

Sequences were sampled at different countries from the east (n = 352), central (n = 53) and southern (n = 1,576) African regions shown in Table 1. The color of branches represents the geographic region from where the subtype C sequences originated, according to the map given in the figure. The boxes highlight the position of the major east African subtype C lineages. The tree was rooted using HIV-1 subtype A1 and D reference sequences (black branches). Horizontal branch lengths are drawn to scale with the bar at the bottom indicating nucleotide substitutions per site.

Figure 2. Geographic distribution of HIV-1 subtype C clades in east Africa.

a) Map of Africa showing the frequency of distinct HIV-1 subtype C clades across the five countries from the east region here studied (Burundi, Ethiopia, Kenya, Uganda and Tanzania). b) Map of Tanzania showing the frequency of distinct HIV-1 subtype C clades across different country regions where patients included in the present study resided (Kagera, Mwanza, Kilimanjaro and Mbeya). The legend for the colors on graphics is shown on the right.

Figure 3. Time-scaled Bayesian MCC tree of the HIV-1 C_EA lineage.

Branches are colored according to the most probable location state of their descendent nodes. The legend for the colors is shown on the left. The state posterior probability is indicated only at key nodes. The boxes highlight the position of the major country-specific sub-clades detected in our study. The median age (with 95% HPD interval in parentheses) of those country-specific sub-clades is shown. Horizontal branch lengths are drawn to scale with the bar at the bottom indicating years. The tree was automatically rooted under the assumption of a relaxed molecular clock.

Figure 4. Spatiotemporal dynamic of HIV-1 C_EA clade dissemination in east Africa.

We provide snapshots of the dispersal pattern for the years 1960, 1965, 1970, 1975, 1980, 1985, 1990 and 2000. Lines between locations represent branches in the Bayesian MCC tree along which location transition occurs. Location circle diameters are proportional to square root of the number of Bayesian MCC branches maintaining a particular location state at each time-point. The white-green color gradient informs the relative age of the transitions (older-recent). The maps are based on satellite pictures made available in Google^™ Earth (http://earth.google.com).

Figure 5. Time-scaled Bayesian MCC tree of major Ethiopian HIV-1 subtype C lineages.

MCC tree was obtained after exclusion of putative C/C′ intrasubtype recombinant sequences. Branches are colored according to the initial clade assignment of each sequence based on ML analysis: C_ET (blue) and C′_ET (red). The PP support and the median age (with 95% HPD interval in parentheses) are indicated only at key nodes. Horizontal branch lengths are drawn to scale with the scale at the bottom indicating years. The tree was automatically rooted under the assumption of a relaxed molecular clock.